Face Generation

In this project, you'll use generative adversarial networks to generate new images of faces.

Get the Data

You'll be using two datasets in this project:

  • MNIST
  • CelebA

Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner.

If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

In [1]:
#data_dir = './data'

# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)
Found mnist Data
Found celeba Data

Explore the Data

MNIST

As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images.

In [2]:
show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')
Out[2]:
<matplotlib.image.AxesImage at 0x7f9bf2edbcf8>

CelebA

The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images.

In [3]:
show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))
Out[3]:
<matplotlib.image.AxesImage at 0x7f9bf2e184a8>

Preprocess the Data

Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28.

The MNIST images are black and white images with a single color channel while the CelebA images have 3 color channels (RGB color channel).

Build the Neural Network

You'll build the components necessary to build a GANs by implementing the following functions below:

  • model_inputs
  • discriminator
  • generator
  • model_loss
  • model_opt
  • train

Check the Version of TensorFlow and Access to GPU

This will check to make sure you have the correct version of TensorFlow and access to a GPU

In [4]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))
TensorFlow Version: 1.2.1
Default GPU Device: /gpu:0

Input

Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders:

  • Real input images placeholder with rank 4 using image_width, image_height, and image_channels.
  • Z input placeholder with rank 2 using z_dim.
  • Learning rate placeholder with rank 0.

Return the placeholders in the following the tuple (tensor of real input images, tensor of z data)

In [5]:
import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    # implemented
    input_real = tf.placeholder(tf.float32, [None, image_width, image_height, image_channels], name='input_real')
    input_z= tf.placeholder(tf.float32, [None, z_dim], name='input_z')
    learning_rate = tf.placeholder(tf.float32, name='learning_rate')
    
    
    return input_real, input_z, learning_rate


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)
ERROR:tensorflow:==================================
Object was never used (type <class 'tensorflow.python.framework.ops.Operation'>):
<tf.Operation 'assert_rank_2/Assert/Assert' type=Assert>
If you want to mark it as used call its "mark_used()" method.
It was originally created here:
['File "/usr/local/lib/python3.5/runpy.py", line 193, in _run_module_as_main\n    "__main__", mod_spec)', 'File "/usr/local/lib/python3.5/runpy.py", line 85, in _run_code\n    exec(code, run_globals)', 'File "/usr/local/lib/python3.5/site-packages/ipykernel_launcher.py", line 16, in <module>\n    app.launch_new_instance()', 'File "/usr/local/lib/python3.5/site-packages/traitlets/config/application.py", line 658, in launch_instance\n    app.start()', 'File "/usr/local/lib/python3.5/site-packages/ipykernel/kernelapp.py", line 477, in start\n    ioloop.IOLoop.instance().start()', 'File "/usr/local/lib/python3.5/site-packages/zmq/eventloop/ioloop.py", line 177, in start\n    super(ZMQIOLoop, self).start()', 'File "/usr/local/lib/python3.5/site-packages/tornado/ioloop.py", line 888, in start\n    handler_func(fd_obj, events)', 'File "/usr/local/lib/python3.5/site-packages/tornado/stack_context.py", line 277, in null_wrapper\n    return fn(*args, **kwargs)', 'File "/usr/local/lib/python3.5/site-packages/zmq/eventloop/zmqstream.py", line 440, in _handle_events\n    self._handle_recv()', 'File "/usr/local/lib/python3.5/site-packages/zmq/eventloop/zmqstream.py", line 472, in _handle_recv\n    self._run_callback(callback, msg)', 'File "/usr/local/lib/python3.5/site-packages/zmq/eventloop/zmqstream.py", line 414, in _run_callback\n    callback(*args, **kwargs)', 'File "/usr/local/lib/python3.5/site-packages/tornado/stack_context.py", line 277, in null_wrapper\n    return fn(*args, **kwargs)', 'File "/usr/local/lib/python3.5/site-packages/ipykernel/kernelbase.py", line 283, in dispatcher\n    return self.dispatch_shell(stream, msg)', 'File "/usr/local/lib/python3.5/site-packages/ipykernel/kernelbase.py", line 235, in dispatch_shell\n    handler(stream, idents, msg)', 'File "/usr/local/lib/python3.5/site-packages/ipykernel/kernelbase.py", line 399, in execute_request\n    user_expressions, allow_stdin)', 'File "/usr/local/lib/python3.5/site-packages/ipykernel/ipkernel.py", line 196, in do_execute\n    res = shell.run_cell(code, store_history=store_history, silent=silent)', 'File "/usr/local/lib/python3.5/site-packages/ipykernel/zmqshell.py", line 533, in run_cell\n    return super(ZMQInteractiveShell, self).run_cell(*args, **kwargs)', 'File "/usr/local/lib/python3.5/site-packages/IPython/core/interactiveshell.py", line 2698, in run_cell\n    interactivity=interactivity, compiler=compiler, result=result)', 'File "/usr/local/lib/python3.5/site-packages/IPython/core/interactiveshell.py", line 2808, in run_ast_nodes\n    if self.run_code(code, result):', 'File "/usr/local/lib/python3.5/site-packages/IPython/core/interactiveshell.py", line 2862, in run_code\n    exec(code_obj, self.user_global_ns, self.user_ns)', 'File "<ipython-input-5-b2f015e924c8>", line 24, in <module>\n    tests.test_model_inputs(model_inputs)', 'File "/output/problem_unittests.py", line 12, in func_wrapper\n    result = func(*args)', 'File "/output/problem_unittests.py", line 68, in test_model_inputs\n    _check_input(learn_rate, [], \'Learning Rate\')', 'File "/output/problem_unittests.py", line 34, in _check_input\n    _assert_tensor_shape(tensor, shape, \'Real Input\')', 'File "/output/problem_unittests.py", line 20, in _assert_tensor_shape\n    assert tf.assert_rank(tensor, len(shape), message=\'{} has wrong rank\'.format(display_name))', 'File "/usr/local/lib/python3.5/site-packages/tensorflow/python/ops/check_ops.py", line 617, in assert_rank\n    dynamic_condition, data, summarize)', 'File "/usr/local/lib/python3.5/site-packages/tensorflow/python/ops/check_ops.py", line 571, in _assert_rank_condition\n    return control_flow_ops.Assert(condition, data, summarize=summarize)', 'File "/usr/local/lib/python3.5/site-packages/tensorflow/python/util/tf_should_use.py", line 170, in wrapped\n    return _add_should_use_warning(fn(*args, **kwargs))', 'File "/usr/local/lib/python3.5/site-packages/tensorflow/python/util/tf_should_use.py", line 139, in _add_should_use_warning\n    wrapped = TFShouldUseWarningWrapper(x)', 'File "/usr/local/lib/python3.5/site-packages/tensorflow/python/util/tf_should_use.py", line 96, in __init__\n    stack = [s.strip() for s in traceback.format_stack()]']
==================================
Tests Passed

Discriminator

Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variables in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the discriminator, tensor logits of the discriminator).

In [27]:
def leaky_rlu(inputs, alpha):
    return tf.maximum(alpha*inputs, inputs)
def discriminator(images, reuse=False):
    """
    Create the discriminator network
    :param images: Tensor of input image(s)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """ 
    def conv(inputs, filters, filter_shape, strides, training, alpha):
        second = tf.layers.conv2d(inputs, filters, filter_shape, strides, padding='same')
        nor_narr_second = tf.layers.batch_normalization(second, training=training)
        return leaky_rlu(nor_narr_second, alpha)
    # implemented
    with tf.variable_scope('discriminator', reuse=reuse):
        #input 28*28*3
        # c1 14*14*3
        c1 = conv(images, 64, 5, 2, True, 0.2)
        # c2 7*7*3
        c2 = conv(c1, 128, 5, 2, True, 0.2)
         # c3 4*4*3
        c3 = conv(c2, 256, 5, 2, True, 0.2)
        
        flatt = tf.reshape(c3, [-1,4*4*256])
        logits = tf.layers.dense(flatt, 1)
        out = tf.sigmoid(logits)
    return out, logits


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)
Tests Passed

Generator

Implement generator to generate an image using z. This function should be able to reuse the variables in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images.

In [29]:
def generator(z, out_channel_dim, is_train=True):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    def transposed_conv(inputs, filters, filter_shape, strides, training, alpha):
        tcon = tf.layers.conv2d_transpose(inputs, filters, filter_shape, strides, padding='same')
        ntcon = tf.layers.batch_normalization(tcon, training=training)
        return leaky_rlu(ntcon, alpha)
    # implemented
    with tf.variable_scope('generator', reuse=(is_train==False)) as scope:
        # d1 4*4*512
        d1 = tf.layers.dense(z, 4*4*512, activation=None, use_bias=False)
        nd1 = tf.reshape(d1, (-1, 4, 4, 512))
        nnd1 = tf.layers.batch_normalization(nd1, training=is_train)
        rlu_nnd1 = leaky_rlu(nnd1, 0.2)

        # c1 7*7*256
        try:
            weights = tf.get_variable('c1_weight', shape=[5, 5, 256, 512],initializer=tf.truncated_normal_initializer(stddev=0.05))
        except ValueError:
            scope.reuse_variables()
            weights = tf.Variable(tf.truncated_normal([5, 5, 256, 512], stddev=0.05))
        
        c1 = tf.nn.conv2d_transpose(rlu_nnd1, weights, [32, 7, 7, 256], strides=[1, 2, 2, 1], padding='SAME', name='c1')
        nc1 = tf.layers.batch_normalization(c1, training=is_train)
        rnc1 = leaky_rlu(nc1, 0.2)
            
        # c2 14*14*128  
        c2 = transposed_conv(rnc1, 128, 5, [2, 2], is_train, 0.2)
         
        # c3 28*28*3 
        c3 = tf.layers.conv2d_transpose(c2, out_channel_dim, 5, [2, 2], padding='same')

        out = tf.tanh(c3)/2
        return out


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)
Tests Passed

Loss

Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented:

  • discriminator(images, reuse=False)
  • generator(z, out_channel_dim, is_train=True)
In [30]:
def model_loss(input_real, input_z, out_channel_dim):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """
    # implemented
    d_real_out,d_real_logits = discriminator(input_real, reuse=False)
    d_fake_out,d_fake_logits = discriminator(generator(input_z, out_channel_dim, is_train=True), reuse=True)
    
    d_loss_real = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_real_logits, labels=tf.ones_like(d_real_out)))
    d_loss_fake = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_fake_logits, labels=tf.zeros_like(d_fake_out)))
    g_loss = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_fake_logits, labels=tf.ones_like(d_fake_out)))

    d_loss = d_loss_real + d_loss_fake

    return d_loss, g_loss


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)
Tests Passed
In [31]:
def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    # Get weights and bias to update
    t_vars = tf.trainable_variables()
    d_vars = [var for var in t_vars if var.name.startswith('discriminator')]
    g_vars = [var for var in t_vars if var.name.startswith('generator')]

    # Optimize
    with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
        d_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(d_loss, var_list=d_vars)
        g_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(g_loss, var_list=g_vars)

    return d_train_opt, g_train_opt


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)
Tests Passed

Optimization

Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation).

Neural Network Training

Show Output

Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training.

In [32]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

Train

Implement train to build and train the GANs. Use the following functions you implemented:

  • model_inputs(image_width, image_height, image_channels, z_dim)
  • model_loss(input_real, input_z, out_channel_dim)
  • model_opt(d_loss, g_loss, learning_rate, beta1)

Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches.

In [33]:
def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode, show_every=100):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """
    # Build Model
    input_real, input_z, lr = model_inputs(data_shape[1], data_shape[2], data_shape[3], z_dim)
    d_loss, g_loss = model_loss(input_real, input_z, data_shape[3])
    d_opt, g_opt = model_opt(d_loss, g_loss, lr, beta1)
    ebatch_count = (data_shape[0]+batch_size-1) // batch_size
    batch_count = ebatch_count * epoch_count
    steps = 0
    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        for epoch_i in range(epoch_count):
            for batch_images in get_batches(batch_size):
                steps += 1
                
                # Train Model
                # Sample random noise for G
                batch_z = np.random.uniform(-1, 1, size=(batch_size, z_dim))

                # Run optimizers
                _ = sess.run(d_opt, feed_dict={input_real: batch_images, input_z: batch_z, lr: learning_rate})
                _ = sess.run(g_opt, feed_dict={input_z: batch_z, input_real: batch_images, lr: learning_rate})

                if steps % show_every == 0 or steps == ebatch_count-1:
                    # At the end of each epoch, get the losses and print them out
                    train_loss_d = d_loss.eval({input_z: batch_z, input_real: batch_images})
                    train_loss_g = g_loss.eval({input_z: batch_z})

                    print("Epoch {}/{}...".format(epoch_i+1, epoch_count),
                          "Batch {}/{}...".format(steps+1, batch_count),
                          "Discriminator Loss: {:.4f}...".format(train_loss_d),
                          "Generator Loss: {:.4f}".format(train_loss_g))

                    show_generator_output(sess, 32, input_z, data_shape[3], data_image_mode)

MNIST

Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0.

In [ ]:
batch_size = 32
z_dim = 100
learning_rate = 0.001
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 2

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)
Epoch 1/2... Batch 101/3750... Discriminator Loss: 1.8946... Generator Loss: 0.6477
Epoch 1/2... Batch 201/3750... Discriminator Loss: 1.1251... Generator Loss: 0.5899
Epoch 1/2... Batch 301/3750... Discriminator Loss: 1.0598... Generator Loss: 0.9726
Epoch 1/2... Batch 401/3750... Discriminator Loss: 1.2232... Generator Loss: 0.7149
Epoch 1/2... Batch 501/3750... Discriminator Loss: 0.9075... Generator Loss: 1.2644
Epoch 1/2... Batch 601/3750... Discriminator Loss: 1.1123... Generator Loss: 1.1094
Epoch 1/2... Batch 701/3750... Discriminator Loss: 1.3885... Generator Loss: 0.3953
Epoch 1/2... Batch 801/3750... Discriminator Loss: 1.2263... Generator Loss: 0.9855
Epoch 1/2... Batch 901/3750... Discriminator Loss: 1.4928... Generator Loss: 0.3652
Epoch 1/2... Batch 1001/3750... Discriminator Loss: 1.3245... Generator Loss: 0.4531
Epoch 1/2... Batch 1101/3750... Discriminator Loss: 2.0945... Generator Loss: 0.1558
Epoch 1/2... Batch 1201/3750... Discriminator Loss: 1.5623... Generator Loss: 0.3664
Epoch 1/2... Batch 1301/3750... Discriminator Loss: 0.6625... Generator Loss: 1.7066
Epoch 1/2... Batch 1401/3750... Discriminator Loss: 0.6009... Generator Loss: 1.6560
Epoch 1/2... Batch 1501/3750... Discriminator Loss: 0.5784... Generator Loss: 1.8931
Epoch 1/2... Batch 1601/3750... Discriminator Loss: 0.6449... Generator Loss: 1.0012
Epoch 1/2... Batch 1701/3750... Discriminator Loss: 0.5158... Generator Loss: 1.3556
Epoch 1/2... Batch 1801/3750... Discriminator Loss: 0.5517... Generator Loss: 1.3203
Epoch 1/2... Batch 1875/3750... Discriminator Loss: 0.5953... Generator Loss: 2.0036
Epoch 2/2... Batch 1901/3750... Discriminator Loss: 0.6277... Generator Loss: 1.8111
Epoch 2/2... Batch 2001/3750... Discriminator Loss: 0.4740... Generator Loss: 1.4948
Epoch 2/2... Batch 2101/3750... Discriminator Loss: 2.3452... Generator Loss: 0.1920
Epoch 2/2... Batch 2201/3750... Discriminator Loss: 0.5319... Generator Loss: 1.3237
Epoch 2/2... Batch 2301/3750... Discriminator Loss: 0.4407... Generator Loss: 2.4822
Epoch 2/2... Batch 2401/3750... Discriminator Loss: 0.9705... Generator Loss: 3.3988
Epoch 2/2... Batch 2501/3750... Discriminator Loss: 1.5887... Generator Loss: 2.5033
Epoch 2/2... Batch 2601/3750... Discriminator Loss: 1.0810... Generator Loss: 0.6708
Epoch 2/2... Batch 2701/3750... Discriminator Loss: 0.6011... Generator Loss: 1.3753
Epoch 2/2... Batch 2801/3750... Discriminator Loss: 1.1957... Generator Loss: 0.5576
Epoch 2/2... Batch 2901/3750... Discriminator Loss: 0.5633... Generator Loss: 1.5547
Epoch 2/2... Batch 3001/3750... Discriminator Loss: 0.3470... Generator Loss: 1.9829
Epoch 2/2... Batch 3101/3750... Discriminator Loss: 1.2823... Generator Loss: 0.5694
Epoch 2/2... Batch 3201/3750... Discriminator Loss: 0.6543... Generator Loss: 1.2958
Epoch 2/2... Batch 3301/3750... Discriminator Loss: 1.0661... Generator Loss: 0.6600
Epoch 2/2... Batch 3401/3750... Discriminator Loss: 0.7863... Generator Loss: 0.9141
Epoch 2/2... Batch 3501/3750... Discriminator Loss: 0.4790... Generator Loss: 1.3808
Epoch 2/2... Batch 3601/3750... Discriminator Loss: 0.6823... Generator Loss: 1.6599
Epoch 2/2... Batch 3701/3750... Discriminator Loss: 1.0853... Generator Loss: 0.7100

CelebA

Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.

In [ ]:
batch_size = 32
z_dim = 100
learning_rate = 0.001
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 1

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)
Epoch 1/1... Batch 101/6332... Discriminator Loss: 1.0861... Generator Loss: 3.9655
Epoch 1/1... Batch 201/6332... Discriminator Loss: 1.6172... Generator Loss: 0.4605
Epoch 1/1... Batch 301/6332... Discriminator Loss: 0.9636... Generator Loss: 0.8337
Epoch 1/1... Batch 401/6332... Discriminator Loss: 0.9907... Generator Loss: 0.6143
Epoch 1/1... Batch 501/6332... Discriminator Loss: 1.0427... Generator Loss: 0.5781
Epoch 1/1... Batch 601/6332... Discriminator Loss: 1.5981... Generator Loss: 0.3578
Epoch 1/1... Batch 701/6332... Discriminator Loss: 1.5001... Generator Loss: 0.4783
Epoch 1/1... Batch 801/6332... Discriminator Loss: 2.0733... Generator Loss: 3.5507
Epoch 1/1... Batch 901/6332... Discriminator Loss: 0.6868... Generator Loss: 1.2697
Epoch 1/1... Batch 1001/6332... Discriminator Loss: 1.0674... Generator Loss: 0.8301
Epoch 1/1... Batch 1101/6332... Discriminator Loss: 0.7990... Generator Loss: 1.8363
Epoch 1/1... Batch 1201/6332... Discriminator Loss: 1.1014... Generator Loss: 1.7943
Epoch 1/1... Batch 1301/6332... Discriminator Loss: 2.0408... Generator Loss: 0.1867
Epoch 1/1... Batch 1401/6332... Discriminator Loss: 1.7637... Generator Loss: 0.3548
Epoch 1/1... Batch 1501/6332... Discriminator Loss: 1.3889... Generator Loss: 0.4459
Epoch 1/1... Batch 1601/6332... Discriminator Loss: 0.4400... Generator Loss: 2.7880
Epoch 1/1... Batch 1701/6332... Discriminator Loss: 0.9420... Generator Loss: 0.5663
Epoch 1/1... Batch 1801/6332... Discriminator Loss: 1.4886... Generator Loss: 0.3874
Epoch 1/1... Batch 1901/6332... Discriminator Loss: 0.8351... Generator Loss: 2.8594
Epoch 1/1... Batch 2001/6332... Discriminator Loss: 1.1241... Generator Loss: 0.5807
Epoch 1/1... Batch 2101/6332... Discriminator Loss: 0.6518... Generator Loss: 1.0051
Epoch 1/1... Batch 2201/6332... Discriminator Loss: 0.4097... Generator Loss: 1.9459
Epoch 1/1... Batch 2301/6332... Discriminator Loss: 0.3222... Generator Loss: 2.4585
Epoch 1/1... Batch 2401/6332... Discriminator Loss: 0.1873... Generator Loss: 2.9580
Epoch 1/1... Batch 2501/6332... Discriminator Loss: 0.3255... Generator Loss: 2.6010
Epoch 1/1... Batch 2601/6332... Discriminator Loss: 0.2835... Generator Loss: 3.1401
Epoch 1/1... Batch 2701/6332... Discriminator Loss: 1.4857... Generator Loss: 3.1930
Epoch 1/1... Batch 2801/6332... Discriminator Loss: 0.6025... Generator Loss: 1.1690
Epoch 1/1... Batch 2901/6332... Discriminator Loss: 0.3008... Generator Loss: 1.8335
Epoch 1/1... Batch 3001/6332... Discriminator Loss: 2.9849... Generator Loss: 5.4523
Epoch 1/1... Batch 3101/6332... Discriminator Loss: 1.1689... Generator Loss: 0.5513
Epoch 1/1... Batch 3201/6332... Discriminator Loss: 0.3991... Generator Loss: 1.4896
Epoch 1/1... Batch 3301/6332... Discriminator Loss: 1.0719... Generator Loss: 4.7434
Epoch 1/1... Batch 3401/6332... Discriminator Loss: 0.1491... Generator Loss: 2.8010
Epoch 1/1... Batch 3501/6332... Discriminator Loss: 0.2798... Generator Loss: 2.2971
Epoch 1/1... Batch 3601/6332... Discriminator Loss: 0.8359... Generator Loss: 0.7781
Epoch 1/1... Batch 3701/6332... Discriminator Loss: 0.6901... Generator Loss: 1.0035
Epoch 1/1... Batch 3801/6332... Discriminator Loss: 0.4882... Generator Loss: 3.7569
Epoch 1/1... Batch 3901/6332... Discriminator Loss: 0.6226... Generator Loss: 2.4586
Epoch 1/1... Batch 4001/6332... Discriminator Loss: 0.4653... Generator Loss: 2.7132
Epoch 1/1... Batch 4101/6332... Discriminator Loss: 0.9870... Generator Loss: 3.8957
Epoch 1/1... Batch 4201/6332... Discriminator Loss: 0.2892... Generator Loss: 2.8240
Epoch 1/1... Batch 4301/6332... Discriminator Loss: 0.4794... Generator Loss: 2.4108
Epoch 1/1... Batch 4401/6332... Discriminator Loss: 1.0017... Generator Loss: 0.7201
Epoch 1/1... Batch 4501/6332... Discriminator Loss: 0.4890... Generator Loss: 6.0841
Epoch 1/1... Batch 4601/6332... Discriminator Loss: 0.4755... Generator Loss: 3.7121
Epoch 1/1... Batch 4701/6332... Discriminator Loss: 0.8507... Generator Loss: 0.9716
Epoch 1/1... Batch 4801/6332... Discriminator Loss: 0.5241... Generator Loss: 3.9738
Epoch 1/1... Batch 4901/6332... Discriminator Loss: 0.1930... Generator Loss: 2.6498
Epoch 1/1... Batch 5001/6332... Discriminator Loss: 2.5563... Generator Loss: 0.2269
Epoch 1/1... Batch 5101/6332... Discriminator Loss: 0.4215... Generator Loss: 2.6044
Epoch 1/1... Batch 5201/6332... Discriminator Loss: 0.6189... Generator Loss: 1.0561
Epoch 1/1... Batch 5301/6332... Discriminator Loss: 0.2292... Generator Loss: 2.1913
Epoch 1/1... Batch 5401/6332... Discriminator Loss: 0.2330... Generator Loss: 2.6550
Epoch 1/1... Batch 5501/6332... Discriminator Loss: 0.3755... Generator Loss: 1.5508

Submitting This Project

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_face_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.